From W3C Wiki

Derived from RRSAgent minutes


See also: IRC log



Arnaud, aaronpk, shanehudson, sandro, elf-pavlik, kevinmarks, wilkie, eprodrom, jasnell, ben_thatmustbeme, cwebber, tantek, hhalpin, james, tsyesika, wseltzer, akuckartz, shepazu, Rob_Sanderson, Shane_, rene
csarven, rhiaro

<trackbot> Date: 03 November 2015

Made dinner, ready at 1 minute to the hour. That is good timing!

<aaronpk> someone is really noisy on the call

<cwebber2> we miss you Zakim !

<azaroth> cwebber2++

<Loqi> cwebber2 has 51 karma

I can scribe

<eprodrom> Sorry, is the code still SOCL?

<Arnaud> no, you need to follow the link from the meeting page

<Arnaud> we can't type it here or it will be public

<scribe> scribeNick: Shane_

<aaronpk> it's in the channel topic O.o

<wilkie> haha it's in the title


<hhalpin> Yep, I put in there just while people we're dialing in - will change it once we get everyone in.

<cwebber2> yeah I got in with no password

I normally need don't need a password but did this time

<Arnaud> PROPOSED: Approval of Minutes of 2015-10-20

<eprodrom> Scanning now


<cwebber2> I think the standard procedure is drop on irc and everyone scrambles to review :)

<eprodrom> +1 looks complete

<tantek> cwebber2++

<cwebber2> +1 on approval

<Arnaud> RESOLVED: Approval of Minutes of 2015-10-20

<Loqi> cwebber2 has 52 karma

Arnaud: Quickly another reminder about face to face meeting, if you haven't done so please indicate your attendance.
... we will soon start working on logistics for the face to face

<tantek> with AnnB and rhiaro !

Arnaud: Tantek entertained a breakout session at TPAC, about 25 people or so at the session

<tantek> rhiaro took minutes

Arnaud: Not much to take away from the working group, but worth reading through the minutes

tantek: The discussion was very lively and had interest for participation in the WG


tantek: Photo there, you can see lots of people there. You can see Tim Berners-Lee attended, it was very animated

<tantek> rhiaro's minutes in IRC:

Arnaud: The primary goal was to advertise the WG and any new members so much the better
... Trying to streamline the discussion on the agenda
... Evan has voluntered to be a co-editor

<jasnell> +1 to Evan being co-editor

Arnaud: If anyone is concerned, let us know

<cwebber2> massive +1 to Evan being co-editor

<azaroth> +1 and Evan++

<hhalpin_> +1

<rene> +1

Evan: My only question is if there is any conflict with being a chair and co-editor?

<cwebber2> I'm also happy to be an editor if that would be helpful, though I think Evan + James is more than sufficient

<wilkie> +1

<hhalpin_> Its unusual but three chairs is also unusual.

Arnaud: If you are willing to put in the time it takes, you are very welcome

<ben_thatmustbeme> +1

<eprodrom> +1

<tantek> +1

<eprodrom> Thanks everyone


<hhalpin_> I would suggest that if you feel there is a conflict of interest between the two roles, you simply tell the other chairs and editor.

Arnaud: I sent an email about the agenda, there was a lot of different proposals regarding AS 2.0 and it would be time consuming to go through them one by one

<jasnell> in the batch of "proposed to accept", let's pull [Activity Streams 2.0] Proposal: Remove @context from all examples - (Ben) out separately.

<azaroth> +1 to separating 233, per email

Arnaud: I took the liberty to bunch together some propsals, proposing to approve all as a batch
... If you have any concerns with any one proposal, say so. No need to justify now
... Please speak up, otherwise we will vote on the rest and hopefully agree these are the proposals, then focus on the others on a more systematic approach

<cwebber2> yay to the idea of resolving many proposals at once :)

<Loqi> woot

<cwebber2> ben_thatmustbeme: heh!

Arnaud: There are proposals from a few people

<ben_thatmustbeme> that one has had a fair number of -1s

<jasnell> The Proposal then is to approve: #220, #218, #225, #226, #227, #213, #229, and #232

Arnaud: Is there any request for removal from that batch?

<Arnaud> PROPOSED: Approve the following proposals: #220, #218, #225, #226, #227, #213, #229, and #232

<cwebber2> +1

<jasnell> +1

<azaroth> +1

<rene> +1


<aaronpk> +1

<tantek> +1

<ben_thatmustbeme> wait, that list is missing 223

<aaronpk> that's correct

<tantek> it's missing 233

<cwebber2> ben_thatmustbeme: that's up for seaprate discussion

<ben_thatmustbeme> okay

<ben_thatmustbeme> +1

<tantek> the remove @context from all examples - postponed to specific discussion after this poll

<ben_thatmustbeme> wasn't sure the way it was nested

Arnaud: Ben, we can look at that one next

<wilkie> +1

<Arnaud> RESOLVED: Approve the following proposals: #220, #218, #225, #226, #227, #213, #229, and #232

<eprodrom> +1

<cwebber2> whooo, talk about getting stuff done!

<tantek> Aside: I think that was a reasonable improvement on the flat/FIFO agenda building.

<Arnaud> [Activity Streams 2.0] Proposal: `mediaType` on Content objects - (James)

<Loqi> Hhalpin made 1 edit to Socialwg

<Loqi> Tantekelik made 1 edit to Socialwg/2015-10-20-minutes

jasnell: This proposal is from Elf, he may be able to explain it better. It is about whether we need object types for Blob, such as image data. We already have Image and Video but that is for describing metadata of the image. The proposal is about the actual data.

Arnaud: Elf is not on today

jasnell: Defer it until he is here

<cwebber2> elves with invisibility cloaks

Arnaud: Unless anyone else wants to speak up?

jasnell: #229 should cover it, but will wait for Elf

Arnaud: There is the @context, do we want to start with that one?

<kevinmarks> just the 9 digit one

Arnaud: Happy to leave it to those who are involved in the discussion, if they would gain from using the teleconference time

<Zakim> ben_thatmustbeme, you wanted to say that I'm okay to waiting for this on non-telcon time

<kevinmarks> POTS via google

ben_thatmustbeme: I'm ok with not wasting time on it. Sounds like annotations group are going to go other direction on it

tantek: The only thing I was going to add is since we have rob and ben on the phone it may be good to have the chance for them to discuss it

Rob: The annotations group originally left them off to make samples clearer, we reversed it on the grounds of developers being likely to cut and paste and make sure their code produces code that looks like the samples. Which is not what we wanted to end up with
... We haven't yet published the reversal

Arnaud: Ben, since the discussion is still going do you want to respond?

<jasnell> fwiw, I'm -1 on #233. having the @context there doesn't hurt and is beneficial

Ben: I'm ok with leaving them there, there have been several -1s so I would be ok leaving them there

<jasnell> paying attention to @context is optional for *some* users

<jasnell> not all

Ben: James has said context is optional in the past, based on other arguments people say it must be there for JSON-LD

<cwebber2> +q

<wilkie> shouldn't it use a default context when absent?

Arnaud: There are people pushing for just JSON. We should not tie this question to the bigger issue

<jasnell> wilkie: it does

Arnaud: We should separate the discussions

<wilkie> jasnell: ok. that makes it seem like less a big deal either way then.

<Arnaud> PROPOSED: Keep @context from all examples - (Ben)

<jasnell> +1 to keeping @context

<azaroth> +1

<rene> +1

<tantek> 0

<ben_thatmustbeme> 0

<cwebber2> I agree taht we need to have that other discussion

<wilkie> +0

<cwebber2> +0

<ben_thatmustbeme> i can live with it

+0 I have no reason not to, not a big fan though

<aaronpk> +0

<melvster> +1

jasnell: The annotations working group have the option of requring JSON-LD things, in this WG we decided to say JSON as in charter but never required JSON-LD

<cwebber2> tantek: I think the points you're raising are important, I agree with Arnaud that we need to have that larger conversation to clarify this

<jasnell> Shane_ that's not me talking

Oh whoops, it is ben then?

<cwebber2> Shane_: it was tantek

oops, sorry!

<rene> looking at the current drafts it seems a lot like JSON-LD is the primary serialization the WG is looking at

Arnaud: It was approved by the working group

tantek: Then we have a difference between a couple of the chairs, we can discuss later

<Arnaud> RESOLVED: Keep @context from all examples - (Ben)

<kevinmarks> at the time we discussed JSON-LD we were told that @context was optional

Arnaud: I can live with a 0 and so will call it resolved

<cwebber2> kevinmarks: it is optional still

<Arnaud> Proposal: Remove "hreflang" -

<cwebber2> kevinmarks: maybe it can be better clarified in the docs, I need to re-look

Arnaud: A proposal to remove href lang

<cwebber2> the context is implied, so

<tantek> right kevinmarks, we only ever resolved that optional JSON-LD was ok. I don't remember ever resolving to require JSONLD in AS2. If we did then we have bigger problems. :/

<azaroth> kevinmarks, cwebber2: +1 to clarifying. 1.2 of AS-core implies it's required, at least to me

<cwebber2> what we resolved in boston iirc

cwebber2: The object wasn't designed to have parity, the justification for removing it hasn't been made

<cwebber2> was that the context was implied

<kevinmarks> right, cwebber2 - the implied @context was the way to map the JSON to JSON-LD for those who want it

Arnaud: Ok, anyone wants to speak to the flip side?

<cwebber2> and that way

<cwebber2> you could have json-ld

<cwebber2> but someone who just has json can just use json

Ben: I can't see how it is useful, yes it is in html but so is a lot of things. We should prove useful not what's not useful. You can't prove a negative

<tantek> +1 on being required to prove that it is useful

<hhalpin_> I think you can prove a 'negative' if its not used by developers in the wild.

<cwebber2> so simultaneously it's both json-ld but not needing to be json-ld

<hhalpin_> That's an empirical question, but we're a bit early here.

<hhalpin_> We could see if hreflang was heavily used in Atom though.


Ben: I didn't see anything very conclusive, I didn't see anything in the language spec/best practises spec that was published

Arnaud: Are you saying you don't buy it as being valid?

<wilkie> if you remove this, how does an extension add it back?

Ben: Not really

<Zakim> azaroth, you wanted to agree with jasnell that it's useful

<jasnell> it's not new...

<tantek> can we postpone it from AS2? whose implementation needs this?

azaroth: We had the same use case with the same outcome, we useful DC language but the result is the same
... The language are not available by content negotiation
... The annotation might be the only source of language, you may not know which source to use

<melvster> tantek: JSON LD generic processing would require @context -- list of adopters:

<wilkie> there needs to be some space for this functionality for this exact thing as extension if it is removed

<kevinmarks> +1 on deleting things not used in practice

hhalpin_: Examples are easy to make. We may be premature in using hreflang, if there are elements that are not used once we go through them then they should be removed from the AS 2.0 spec or the vocab


<tantek> melvster: JSON LD generic processing is not a requirement for any our WG deliverables

Arnaud: If there are certain features that no one wants to implement then we will delete them

<tantek> therefore it cannot assert requirements on any specs


<hhalpin_> That being said, making the spec as simple as possible in the beginning (rather than as complex as possible, as we get with some specs) helps you get implementers for CR!

<Arnaud> Proposal: keep "hreflang" -, revisit at CR

Ben: I'm ok waiting till CR

<cwebber2> +1 to revisiting at CR

<azaroth> heh

<hhalpin_> Marking it 'at risk' makes sense

<rene> +0

+1 to at risk

<hhalpin_> +1

<melvster> +0

<Arnaud> PROPOSED: keep "hreflang" -, marking it "AT RISK"


<azaroth> +1 to keep, at risk

<cwebber2> +1 to keep and put it at risk

<ben_thatmustbeme> +1

<cwebber2> that seems reasonable to me

<aaronpk> +1

<melvster> +0

<wilkie> if we remove it, and yeah maybe it shouldn't be in core, but it's too hard someone to put back in, then that'd be bad for international implementations and bilingual usages.

<jasnell> +1 to keeping, -0 to marking at risk


<tantek> +0 would prefer to drop early, but ok with at-risk to keep things moving forward

<wilkie> +1 keep, 0 to mark at risk

Arnaud: It allows you to drop things without cycling through the whole discussion again

<Arnaud> RESOLVED: keep "hreflang" -, marking it "AT RISK"

<Arnaud> Proposal: Object partOf Collection -

<wilkie> I only want it because it makes it more obvious there can be multiple representations of objects :)

Arnaud: We don't have Amy on the call today

jasnell: Previously we had a memberOf property, that was removed as there wasn't any interest at the time. When we revisited the paging model, part of that was indicating the model of the current page.

Arnaud: So it is two-fold, should we have this information back should we piggy back or bring old name back

<kevinmarks> how is this different from using categories/tags?

jasnell: There are memberOf in other vocabs, so could use one of those as extension

Arnaud: It sounds like she wants the relationship but doesn't care about the name

<jasnell> tantek: AS1 did not have a memberOf property at all

Rob: It would be valuable, regarding the memberOf and other vocabularies, my understanding is that AS has long since decided to create its own vocabulary regardless of other exisiting. So it shouldn't affect us here.

<Zakim> tantek, you wanted to ask which one is closer to AS1/JSON if either?

Rob: With my annotations WG head on, this would be very useful to us

<kevinmarks> in AS1 it was implicit

<eprodrom> unfortunately I have to drop off

<kevinmarks> as there were many streams

<eprodrom> Will read the minutes

Tantek: I was wondering if there is any precedant in AS 1 that could help our decision

jasnell: Nope

<jasnell> in my opinion, this is easily something that can be done via an extension

<rene> if we reject the proposal with the reason that other vocabularies can be used, I would suggest to keep a hint somewhere in the spec how this specific case should be treated with an extension

<jasnell> if it's absolutely necessary

<azaroth> -1 to tagging/categories

<tantek> that was kevinmarks speaking about tagging/categories

jasnell: Folks can do it with tags and categories but this is more about the current properties

Arnaud: There is no backwards compatibility, the question is do we add it or not
... It can be done as an extension and sounds like the annotations working group will do that

<ben_thatmustbeme> +1 to Defer for Amy

<Arnaud> Proposal: Type for object of an Offer - (Evan

Arnaud: I don't want to close it without Amy being here. Let's move on

<cwebber2> Arnaud: evan had to drop off

<wilkie> he left I think

<cwebber2> time to defer!

Arnaud: I missed the fact Evan had dropped off. He has a bunch of issues
... Evan and Elf next week

<Arnaud> Media type discussions -

<tantek> what did AS1/JSON use for media type?

jasnell: The short summary is that currently the AS spec defines the slot, the argument is the build upon the JSON-LD type with a profile parametre

<jasnell> application/ld+json; profile=""

<melvster> seeAlso:

<rene> There seems no consensus on that we are actually building on JSON-LD

<tantek> rene - correct, no consensus that we are building on JSONLD

<tantek> rene - the charter only dictates JSON for the Social Syntax

<hhalpin_> +1 jasnell

<cwebber2> +q

jasnell: We are not supporting all JSON-LD mechanisms, we are not requiring all implementers to use it. I think using JSON-LD is the correct way to do it. There is a discussion on going

<Zakim> azaroth, you wanted to note discussion with IETF last week

<tantek> -1 on JSONLD media type

<jasnell> I have offered the following compromise:

Rob: At TPAC we had a discussion, at IETF meeting there has just been a discussion about having media types at the same weight as the regular media types
... If that was a concern then the annotations group will be pushing them on it, it should be possible
... We considered that we also have a bunch of restrictions, the profile was sufficent to add in the requirements in a machine processable and normative fashion. You must do JSON-LD and use this frame, can't say that normatively of course.
... Not to push this group one way or another, but worth considering

cwebber2: I think this ties in partly with what tantek was discussing about requiring @context, we agreed that there would be an implied @context with JSON-LD under the hood.

<tantek> with *optional* JSON-LD under the hood

<melvster> under the hood == out of band

cwebber2: If someone with a scripting language is just hacking away, they will know it will be provided without having to worry about it. Keeping it simple.

<jasnell> example: {"@context": "", "@type": "Place", "coordinates": [[1,2],[3,4]]} is valid Activity Streams 2.0 but is not valid JSON-LD because JSON-LD does not support GeoJSON lists of lists

cwebber2: It seems weird as the core doesn't do much without the core vocab, the way I understood before is that we would deliver it based on having a mime type. If there are other ways to express it, technically it was always possible.
... I think it is useful to have some way to say to teh users that there is a simple document, it is technically JSON-LD but no need to overcomplicate it

<tantek> cwebber2++ this is exactly what I was distinguishing between this WGs approach and AnnotationsWG

<Loqi> cwebber2 has 53 karma

jasnell: A valid AS 2.0 document is not a valid JSON-LD, it allows things to be used that are not valid. It does say there is additional processing required for some features
... Advertising that as a JSON-LD document is a bit of a lie, because it is not
... There is enough differences between the two that warrants AS having its own media type
... Okay we can address it with a profile but that is optional for implementors to support

<cwebber2> jasnell: manu was telling me there was some likely possibility that the nested lists issue might be handled

jasnell: The JSON-LD spec only says what a implementation SHOULD do

<cwebber2> in json-ld

<cwebber2> in some way

<hhalpin_> Also, rss and atom had their own media-types, so there is precedent

<tantek> hhalpin_ indeed

<rene> if it is true that a valid AS 2.0 document doesn't have to be a valid JSON-LD document, than the profile is no solution

Arnaud: The JSON-LD portion appears to be controversal, some people want more linked data. This is loosing sight of the fact that the spec was written for two possible views of the world, not force one view on everybody

<cwebber2> jasnell: is there any blocker from the "it's technically a compacted json-ld doc" other than the geojson example

<cwebber2> ?

Arnaud: The pressure I see on both sides are pushing for one view or another, the current draft is trying to carve a compromise

<Zakim> azaroth, you wanted to ask if *every* schema should register a media type?

Arnaud: It would be wise to bring up things if they are required to be tuned, try not to get into religious debate about linked data or not

<tantek> it sounds like james's proposed media type allows for the two possible views of the world, not force one view on everybody

<rene> I wonder whether it is necessary to allow AS 2.0 documents that look like JSON-LD but are not valid due to subtle differences

Rob: Given was James' said, it seems justified that a different media type be used, but I would propose that any features that are not JSON-LD compatible will be called out so that people know what would be lost

<rene> +1 to rob

<cwebber2> +1 to having the mimetype and also saying it SHOULD be considered a profile

<azaroth> +1 to jasnell suggestion of equivalent profile

jasnell: For leaving in the application activity + json activity type, but should treat with profile as being equivalent. If they don't care then it is fine.

<cwebber2> because then implementers can ignore that

<cwebber2> but technically, it's true

<cwebber2> I think it's a good route

<azaroth> Propose that profile doc might be good place for listing non-round-trippable content?

Arnaud: We will not close this issue yet, hoping we are starting to see the light

<ben_thatmustbeme> perhaps a note of how to convert AS2 to JSON-LD ?

Arnaud: There are two more items on the agenda, will use the last minute to let cwebber2 make an announcement

<rene> if there are two worlds, why not express that in using different media types?

cwebber2: I have a very short announcement which is I just posted to the list an implementation of AS 2.0 that tries to implement as simplified so that it uses JSON-LD behind the hood

<Zakim> tantek, you wanted to remind folks about IndieWebCamp MIT Nov 7-8! please sign-up on the Guest List if you want to come (no charge, no need to be

tantek: Just a quick announcement, this week coming up is IWC MIT 7th 8th all day, posted a URL into IRC. Can sign up by adding to wiki page, no charge. Welcome to bring any friends, don't need to be a member of WG

Arnaud: Let's close on this, thank you all for joining

<jasnell> I will have an updated WD ready to review by next week

<wilkie> thanks

<azaroth> Thanks all!

Arnaud: Evan should be chairing on Nov 10th

<Loqi> I added a countdown for 11/10 12:00am (#5735)

<cwebber2> thank you Shane_ !

<cwebber2> Shane_++

<tantek> Shane_++ for scribing!!

<Loqi> Shane_ has 1 karma

<wilkie> Shane++

<cwebber2> Arnaud++ for chairing

<Arnaud> trackbot, end meeting

<ben_thatmustbeme> Shane++

<Loqi> Shane_ has 2 karma

<Loqi> Shane has 3 karma

<Loqi> Arnaud has 27 karma

<Loqi> Shane has 4 karma

Sorry about the slight name mismatches heh

<cwebber2> btw jasnell

Summary of Action Items

[End of minutes]